Adam Khoo: 80% of My Portfolio is in THIS

ELI5 / TLDR

Adam Khoo is a Singaporean investing teacher with a portfolio he puts north of $20M. His pitch here: the AI rally is real but parts of it are overpriced, so don’t bet the farm on it. He keeps roughly 40% of his stocks touching AI and 60% in unloved, profitable, non-AI businesses — so that when the music stops and money rotates out of chips, he’s already holding what it rotates into. Buy great businesses only when they trade below what they’re worth, own at least 8–10 of them, and stop trying to predict the market. Also: a chunk of this is a runway to a July webinar and an app called Stock Oracle.

The Full Story

Inflation is the reason you can’t sit in cash

Khoo’s opening argument is that the riskiest financial move is not investing. Cash quietly loses 2–3% of its purchasing power a year to inflation. Compound that and you lose roughly a fifth of your buying power per decade. His framing throughout is that investing only looks risky the way swimming looks risky to someone who never learned — dangerous until you’ve taken the lessons, then merely useful.

He offers a back-of-envelope number for “how much is enough”: annual expenses × 25. A Singapore family spending $10k a month needs about $3M; a single person on $2,500 a month needs about $750k. The catch he flags is inflation again — that target is in today’s dollars, and in 20 years the same lifestyle costs roughly double.

The market isn’t a casino, it’s a supermarket

This is the line he keeps returning to. A gambler tries to guess where a stock goes next week. An investor treats the market as a shop full of businesses on sale, of which — by his count — fewer than 1% are genuinely high-quality. Buy that 1% when it’s discounted and, in his words:

When you look at a market from that perspective, you are no longer a gambler. You become the house. And the house always wins over time.

His two filters are lifted straight from Warren Buffett. First, the economic moat — a durable competitive advantage that lets a company keep growing for 10–20 years, not just this quarter. Second, intrinsic value — what the business is actually worth, which he estimates with a discounted free-cash-flow model run over a conservative 20-year horizon. The discipline is simple: only buy a great company when it trades below that value.

A great business can be a bad investment if you pay too much for it. But a great business is a great investment if you can buy it for less than what it is worth.

He illustrates with property. Overpay $2M for a flat worth $1.5M and you’ve made a bad investment in a nice apartment. The difference, he says, is that desperate, below-value sellers are rare in property but routine in stocks — because every war scare, rate-hike fear or Trump tweet panics the market into dumping good companies 20–50% below value. That panic is where he claims he got rich. The warning attached: cheap lousy companies get cheaper and never recover; only cheap great companies reliably bounce back.

The portfolio: ride the AI wave, but pre-position for its end

The headline number. Khoo splits his stock portfolio into three buckets:

~20% pure AI infrastructure — semiconductors and the like, companies whose whole business is building AI.
~20% AI-adjacent — businesses with real AI exposure but other engines too. Meta (ads + social), Amazon (ads, logistics, e-commerce). If AI cools, these still earn.
~60% little-to-no AI — solid businesses such as specialty insurer Arch Capital (ACGL) that simply make money.

So 40% touches AI, 60% mostly doesn’t. The logic is that he can’t predict when the AI bull market ends — “could be today, could be next year, could be 3 years from now” — only that it will. When it does, the overpriced AI names correct hardest (he expects 50%+), and the money flows into the ignored non-AI names he’s been quietly accumulating. He calls these the “anti-bubble stocks” and reaches for the dot-com parallel: in 1999 the McDonald’s, Walmarts and Boeings of the world were unloved and falling daily while dot-coms soared; in 2000 the dot-coms dropped 80–90% and the money rotated back into the boring profitable names.

He’s careful not to dismiss AI as hype — he uses it daily and thinks it’s genuinely lifting corporate margins. On valuation he leans on the PEG ratio (price-to-earnings divided by earnings growth) rather than raw P/E, which ignores how fast profits are growing. The S&P’s forward P/E of ~21 is above its 5- and 10-year averages, but its PEG of about 1 reads as fairly priced, not bubbly. Nvidia specifically he considers reasonably priced and is still holding — it sells “entire AI factories” and a software ecosystem, not just chips, which he thinks makes its growth more durable than a pure memory-chip maker’s.

The mechanical rules, and when he sells

Khoo is insistent that prediction is a waste of time — “whoever tells you they can predict the market is a liar or an idiot.” His system is deliberately mechanical: identify great companies, know their intrinsic value, add when price falls below it, hold at least 8–10 names to spread key-man and single-stock risk, and buy in slowly at technical support levels rather than all at once.

Selling splits by intent. A trade gets a stop loss and a profit target up front (lose 1R, or take 2R). An investment he sells only when the business deteriorates — losing its moat, growth fading permanently — or when price runs more than 100% above intrinsic value, at which point he scales out. He says he’s been doing exactly that with his most overpriced AI names.

The income engine and the options sideline

A separate ~S$12M portfolio throws off about S$700k a year in passive income, tax-free in Singapore. It’s a 6% blended yield across REITs, investment-grade and high-yield bonds, listed and unlisted private credit, and dividend stocks (heavily the three Singapore banks).

On options, his core move is selling cash-secured puts on solid, undervalued companies — which he frames as selling insurance to nervous fund managers and collecting the premium. Take Microsoft at $400 with an estimated value of $550: he sells a put at a $350 strike, collects (say) $10. If the stock stays above $350 he keeps the premium for nothing; if it drops below, he’s obliged to buy at $350 — net $340 after premium — a price he’s happy with anyway. He calls it win-win, though the honest framing is that he’s paid to take downside risk on stocks he already wants. He also mentions a “bullish synthetic spread” to control 100 shares for a quarter of the capital — saved, naturally, for the webinar.

The advice for everyone else

His most useful blanket line: 90% of people have no interest in analyzing businesses and should just buy an index ETF and dollar-cost-average into it, compounding at 8–11% over time. Only the 10% willing to do the work should pick stocks. And the oldest rule of all, which he credits for his own wealth: always spend less than you earn, regardless of income.

Key Takeaways

The real risk isn’t market volatility, it’s holding cash and letting inflation compound away ~20% of purchasing power per decade.
“Financial freedom” target ≈ annual expenses × 25 — then inflate it for the year you’ll actually retire.
Two filters only: a wide, durable economic moat (sustainable 10–20 years), and a price below intrinsic value (DCF over a 20-year horizon).
A great business bought too expensive is a bad investment; the discount is the whole game.
PEG beats raw P/E for judging an index — the S&P’s ~21 P/E looks fairer once you account for earnings growth (PEG ≈ 1).
Barbell the AI theme: ~40% AI/AI-adjacent, ~60% unloved profitable non-AI names that catch the rotation when AI corrects.
Don’t predict timing; pre-position. The AI top is unknowable, so own what money flows into afterward.
Diversify across at least 8–10 quality names to neutralize key-man and single-stock risk; scale in at support, don’t lump-sum.
Sell an investment only on business deterioration or when price runs >100% over intrinsic value.
For 90% of people: index ETF + dollar-cost averaging beats stock-picking. And spend less than you earn, always.
Income layer: a diversified ~6% yield across REITs, bonds, private credit and dividend stocks can fund a life passively.
Selling cash-secured puts = getting paid to buy stocks you already want at a lower price.

Claude’s Take

The investing philosophy here is sound and largely uncontroversial — it’s Buffett and Peter Lynch repackaged in clean, teachable language, and the barbell framing (don’t predict the AI top, just own both sides of the rotation) is a genuinely sensible way to hold a strong opinion without betting everything on its timing. Khoo is good at analogies and honest about the one thing most gurus won’t admit: that nobody can predict the market.

That said, this is a sales funnel wearing an interview’s clothes. Nearly every concept lands on “…which I’ll show you in the July anti-bubble mania webinar” or “…which Stock Oracle does in three minutes.” The app conveniently scores moats 1–10 and spits out intrinsic value — outputs that depend entirely on assumptions (growth rate, discount rate, 20-year cash flows) that the video never opens up. Treat the moat score and the intrinsic-value number as one analyst’s opinion dressed as a fact, not gospel.

A few claims deserve a raised eyebrow: the unverifiable $20M and S$12M portfolio figures, the “I got rich buying panics” narrative (survivorship-friendly), and the cash-secured-put pitch framed as “win-win, free money” — it isn’t free; you’re being paid to absorb real downside, and in a genuine crash you buy a falling knife at the strike. The 6% income yield leaning on high-yield bonds and private credit also carries more risk than the relaxed tone implies. None of it is dishonest, but the calibration tilts promotional. Score: 6 — solid, well-explained fundamentals, marked down for the persistent webinar/app upsell and the unaudited self-reported numbers.